Combined experimental and computational investigation of the mechanism of nickel-catalyzed three-component addition processes

Article abstract of DOI:10.1021/om049471a

The mechanism of nickel-catalyzed couplings of an enone, alkyne, and organozinc has been studied. Adducts of the substrates with nickel(0) have been isolated and characterized, and their reactivity was examined. A potential intermediate was demonstrated to not be kinetically competent in catalytic cyclizations. A computational approach employing the B3LYP density functional method and the 6-31G(d) basis set was used to examine mechanistic possibilities that were consistent with experimental observations, and a modified mechanism for the catalytic cyclizations was formulated. The newly proposed mechanism involves production of an active catalyst that involves a novel interaction between Ni(0) and dimethylzinc.

Full text of DOI:10.1021/om049471a

4636
Organometallics 2004, 23, 4636-4646
Com bin ed Exp er im en ta l a n d Com p u ta tion a l
In vestiga tion of th e Mech a n ism of Nick el-Ca ta lyzed
Th r ee-Com p on en t Ad d ition P r ocesses
Hrant P. Hratchian,† Sanjoy K. Chowdhury, V´ıctor M. Gutie´rrez-Garc´ıa,
Kande K. D. Amarasinghe, Mary J ane Heeg,‡ H. Bernhard Schlegel,*,† and
J ohn Montgomery*
Department of Chemistry, Wayne State University, Detroit, Michigan 48202-3489
Received J uly 14, 2004
The mechanism of nickel-catalyzed couplings of an enone, alkyne, and organozinc has
been studied. Adducts of the substrates with nickel(0) have been isolated and characterized,
and their reactivity was examined. A potential intermediate was demonstrated to not be
kinetically competent in catalytic cyclizations. A computational approach employing the
B3LYP density functional method and the 6-31G(d) basis set was used to examine
mechanistic possibilities that were consistent with experimental observations, and a modified
mechanism for the catalytic cyclizations was formulated. The newly proposed mechanism
involves production of an active catalyst that involves a novel interaction between Ni(0)
and dimethylzinc.
In tr od u ction
ied processes that involve metallacycles formed by the
oxidative cyclization mechanism.
Nickel metallacycles have been known for decades,
and their methods of preparation include oxidative
cyclization of two π components,1 oxidative addition into
the σ bond of a cyclic precursor,2 addition of a bis-
metalated precursor to a nickel dihalide,3 and ortho
metalation.4 The route involving oxidative cyclization
of two π components is of particular synthetic signifi-
cance, since the formation of metallacycles by such a
sequence is believed to be a critical mechanistic step in
a number of important catalytic processes. For instance,
the dimerization and oligomerization of butadienes,1a,5
the [4 + 2] cycloaddition of alkynes and dienes,6 and
the cyclotrimerization of alkynes7 are extensively stud-
In studies from our own laboratories and others, a
number of nickel-catalyzed processes have been studied
that involve three-component couplings of two π com-
ponents with a main-group organometallic.1c,8 Whereas
a number of mechanisms have been considered for the
different variants studied, a metallacycle-based mech-
anism is consistent with most of the experimental
observations that have resulted from this program
(Scheme 1). The formation of the bis-π-complex 1,
(5) (a) Benn, R.; Bu¨ssemeier, B.; Holle, S.; J olly, P. W.; Mynott, R.;
Tkatchenko, I.; Wilke, G. J . Organomet. Chem. 1985, 279, 63. (b) Keim,
W. Angew. Chem., Int. Ed. Engl. 1990, 29, 235. (c) Wender, P. A.; Ihle,
N. C. Tetrahedron Lett. 1987, 28, 2451. (d) Wender, P. A.; Snapper,
M. L. Tetrahedron Lett. 1987, 28, 2221. (e) Wender, P. A.; Ihle, N. C.;
Correia, C. R. D. J . Am. Chem. Soc. 1988, 110, 5904. (f) Wender, P.
A.; Tebbe, M. J . Synthesis 1991, 1089. (g) Wender, P. A.; Nuss, J . M.;
Smith, D. B.; Sua´rez-Sobrino, A.; Vagberg, J .; Decosta, D.; Bordner, J .
J . Org. Chem. 1997, 62, 4908. (h) Wender, P. A.; Ihle, N. C. J . Am.
Chem. Soc. 1986, 108, 4678.
(6) (a) Wender, P. A.; J enkins, T. E. J . Am. Chem. Soc. 1989, 111,
6432. (b) Wender, P. A.; Smith, T. E. J . Org. Chem. 1995, 60, 2962. (c)
Wender, P. A.; J enkins, T. E.; Suzuki, S. J . Am. Chem. Soc. 1995, 117,
1843. (d) Wender, P. A.; Smith, T. E. J . Org. Chem. 1996, 61, 824. (e)
Wender, P. A.; Smith, T. E. Tetrahedron 1998, 54, 1255.
(7) (a) Reppe, W.; Schlichting, O.; Klager, K.; Toepel, T. J ustus
Liebigs Ann. Chem. 1948, 560, 1. (b) Lautens, M.; Klute, W.; Tam, W.
Chem. Rev. 1996, 96, 49. (c) Bhatarah, P.; Smith, E. H. J . Chem. Soc.,
Perkin Trans. 1 1992, 2163. (d) Sato, Y.; Nishimata, T.; Mori, M. J .
Org. Chem. 1994, 59, 6133. (e) Ikeda, S.; Watanabe, H.; Sato, Y. J .
Org. Chem. 1998, 63, 7026. (f) Shanmugasundaram, M.; Wu, M. S.;
J eganmohan, M.; Huang, C.-W.; Cheng, C.-H. J . Org. Chem. 2002, 67,
7724. See also: (g) Louie, J .; Gibby, J . E.; Farnworth, M. V.; Tekavec,
T. N. J . Am. Chem. Soc. 2002, 124, 15188.
(8) For representative examples, see: (a) Montgomery, J . Acc. Chem.
Res. 2000, 33, 467. (b) Ikeda, S. Acc. Chem. Res. 2000, 33, 511. (c) Sato,
Y.; Takimoto, M.; Mori, M. J . Am. Chem. Soc. 2000, 122, 1624. (d)
Kimura, M.; Ezoe, A.; Shibata, K.; Tamaru, Y. J . Am. Chem. Soc. 1998,
120, 4033. (e) Kimura, M.; Fujimatsu, H.; Ezoe, A.; Shibata, K.;
Shimizu, M.; Matsumoto, S.; Tamaru, Y. Angew. Chem., Int. Ed. 1999,
38, 397. (f) Ezoe, A.; Kimura, M.; Inoue, T.; Mori, M.; Tamaru, Y.
Angew. Chem., Int. Ed. 2002, 41, 2784. (g) Miller, K. M.; Huang, W.-
S.; J amison, T. F. J . Am. Chem. Soc. 2003, 125, 3442. (h) Patel, S. J .;
J amison, T. F. Angew. Chem., Int. Ed. 2003, 42, 1364. (i) Montgomery,
J . Angew. Chem., Int. Ed. 2004, 43, 3890.
* To whom correspondence should be addressed. E-mail: hbs@
chem.wayne.edu (H.B.S.); jwm@chem.wayne.edu (J .M.).
† Institute for Scientific Computing, Wayne State University.
‡ To whom correspondence regarding X-ray crystallographic deter-
minations should be directed.
(1) (a) Wilke, G. Angew. Chem., Int. Ed. Engl. 1988, 27, 185. (b)
Smith, A. K. In Comprehensive Organometallic Chemistry II; Pud-
dephatt, R. J ., Ed.; Pergamon Press: New York, 1995; Vol. 9, p 29. (c)
Montgomery, J . Organometallic Complexes of Nickel. In Science of
Synthesis; Trost, B. M., Lautens, M., Eds.; Thieme: Stuttgart, Ger-
many, 2001; Vol. 1, pp 11-62. (d) Kaschube, W.; Schro¨der, W.;
Po¨rschke, K. R.; Angermund, K.; Kru¨ger, C. J . Organomet. Chem. 1990,
389, 399. (e) Grubbs, R. H.; Miyashita, A. J . Am. Chem. Soc. 1978,
100, 1300. (f) McKinney, R. J .; Thorn, D. L.; Hoffman, R.; Stockis, A.
J . Am. Chem. Soc. 1981, 103, 2595. (g) Bu¨ch, H. M.; Binger, P.; Benn,
R.; Rufinska, A. Organometallics 1987, 6, 1130. (h) Koo, K.; Hillhouse,
G. L. Organometallics 1998, 17, 2924. (i) Maekawa, M.; Munakata,
M.; Kuroda-Sowa, T.; Hachiya, K. Inorg. Chim. Acta 1995, 230, 249.
(2) (a) Echavarren, A. M.; Castan˜o, A. M. Advances in Metal-Organic
Chemistry; Liebeskind, L. S., Ed.; J AI Press: London, 1998; Vol. 6, p
1. (b) Bercot, E. A.; Rovis, T. J . Am. Chem. Soc. 2002, 124, 174. (c)
Deming, T. J . J . Am. Chem. Soc. 1998, 120, 4240. (d) Deming, T. J .;
Curtin, S. A. J . Am. Chem. Soc. 2000, 122, 5710.
(3) (a) Grubbs, R. H.; Miyashita, A.; Liu, M.; Burk, P. L. J . Am.
Chem. Soc. 1977, 99, 3863. (b) Grubbs, R. H.; Miyashita, A.; Liu, M.;
Burk, P. L. J . Am. Chem. Soc. 1978, 100, 2418.
(4) Carmona, E.; Gutie´rrez-Puebla, E.; Mar´ın, J . M.; Monge, A.;
Paneque, M.; Poveda, M. L.; Ruiz, C. J . Am. Chem. Soc. 1989, 111,
2883.
10.1021/om049471a CCC: $27.50 © 2004 American Chemical Society
Publication on Web 09/02/2004

Combined Experimental and Computational Investigation
Organometallics, Vol. 23, No. 20, 2004 4637
Sch em e 1
Sch em e 2
followed by oxidative cyclization, would afford metalla-
cycle 2. Transmetalation of the main-group organome-
tallic to 2 would afford intermediate 3, and reductive
elimination of this structure would afford the observed
three-component coupling product 4. The π components
that participate in processes of this type include enones,
unsaturated acyloxazolidinones, nitroalkenes, alkyli-
dene malonates, alkynes, dienes, allenes, aldehydes, and
ketones. The organometallics that have been employed
in various applications include organozincs, organoalu-
minums, organoboranes, alkenylzirconiums, and acety-
lenic tins, as well as metal hydrides of silicon and
aluminum. Given the considerable attention that has
recently been devoted to this class of reactions, we have
attempted to provide insight into the mechanism of
these processes by a combined experimental and theo-
retical approach. In this work, we have isolated, char-
acterized, and examined the reactivity of several orga-
nometallic species that may be involved in the processes,
and we have carried out an initial investigation of
asymmetric processes utilizing a ligand class that is
analogous to those that provide well-behaved nickel
complexes. We have used these experimental observa-
tions about structure and reactivity as a starting point
for a hybrid density functional theory study of the
reaction mechanism of the three-component nickel-
catalyzed addition of an enone, alkyne, and organozinc.
The combination of these approaches allows us to
evaluate the previously proposed mechanism for this
reaction class and to propose a modified description.
This interactive and collaborative use of experiment and
theory allows a detailed analysis of mechanistic issues
that could not be resolved by either individual approach.
Sch em e 3
strates of interest could provide a very useful starting
point for additional study.
With this goal in mind, alkynyl enal 13a was treated
with a stoichiometric quantity of Ni(COD)2 and tmeda
(Scheme 3, Figure 1).9a A deep red solution resulted, and
upon concentration, filtration, and recrystallization with
hexane/THF, a red crystalline solid was obtained.
Single-crystal X-ray analysis of the crystalline solid 14a
was carried out, and the molecular structure is the η1
O-bound tautomer of the metallacycle proposed in
catalytic cyclizations. A square-planar geometry is
observed, and the Ni-O-C bond angle of 125° clearly
demonstrates the η1 nature of the enolate. The metal-
lacycle was further characterized by IR, 1H and 13C
NMR, and elemental analysis, and NMR assignments
were rigorously made utilizing 2D techniques. The η1
O-bound solution structure was confirmed by the spec-
troscopic data. The O-bound structure is unusual, since
most classes of late-metal enolates exist as the C-bound
tautomer. The extensive studies from Bergman and
Heathcock with a variety of mid- and late-metal enolates
clearly document this trend, and enolates of nickel were
reported in their study to be η1 C-bound.10 Significantly,
Str u ctu r a l Stu d ies
Whereas many variants of three-component nickel-
catalyzed couplings have been developed, we viewed the
addition of an enone, alkyne, and organozinc (i.e.
generation of products 7 and 8) as a particularly
relevant place to begin our studies (Scheme 2). Not only
had the synthetic utility of the process been well studied
by Ikeda and by us, but the metallacycle derived from
oxidative cyclization of an enone and alkyne had also
been proposed as a key common intermediate in a very
diverse range of synthetic procedures.8a,b,9 For instance,
in addition to the production of compounds 7 and 8 in
organozinc-mediated cyclizations, additional reactivity
manifolds, including [2 + 2 + 2] cycloadditions (products
9 and 10), [3 + 2] alkylative cycloadditions (product 11),
and [2 + 1] oxidative cycloadditions (product 12) had
been demonstrated in processes that likely involved the
intermediacy of metallacycle 6. Therefore, we envisioned
that some knowledge of the precise structural interac-
tions between a Ni(0) catalyst and the organic sub-
(9) (a) Amarasinghe, K. K. D.; Chowdhury, S. K.; Heeg, M. J .;
Montgomery, J . Organometallics 2001, 20, 370. (b) Chowdhury, S. K.;
Amarasinghe, K. K. D.; Heeg, M. J .; Montgomery, J . J . Am. Chem.
Soc. 2000, 122, 6775. (c) Seo, J .; Chui, H. M. P.; Heeg, M. J .;
Montgomery, J . J . Am. Chem. Soc. 1999, 121, 476. (d) Mahandru, G.
M.; Skauge, A. R. L.; Chowdhury, S. K.; Amarasinghe, K. K. D.; Heeg,
M. J .; Montgomery, J . J . Am. Chem. Soc. 2003, 125, 13481. (e) Ikeda,
S.; Kondo, K.; Sato, Y. J . Org. Chem. 1996, 61, 8248. (f) Rayabarapu,
D. K.; Sambaiah, T.; Cheng, C.-H. Angew. Chem., Int. Ed. 2001, 40,
1286. (g) Rayabarapu, D. K.; Cheng, C.-H. Chem. Eur. J . 2003, 9, 3164.

4638 Organometallics, Vol. 23, No. 20, 2004
Hratchian et al.
F igu r e 3. X-ray crystal structure of complex 16. Selected
bond lengths (Å) and angles (deg): Ni-C49 ) 2.015(6), Ni-
C50 ) 2.004(6), Ni-C54 ) 2.017(6), Ni-C55 ) 2.051(7);
C50-Ni-C54 ) 97.6(3).
F igu r e 1. X-ray crystal structure of metallacycle 14a .
Selected bond lengths (Å) and angles (deg): Ni-O )
1.852(6), O-C1 ) 1.316(10), C1-C2 ) 1.298(11), C2-C3
) 1.508(11), C7-C8 ) 1.357(10) C8-Ni ) 1.897(7); O-Ni-
C8 ) 91.3(3), O-Ni-N2 ) 88.9(3), N2-Ni-N1 ) 85.6(3),
N1-Ni-C8 ) 94.3(3), Ni-O-C1 ) 124.7(5), O-C1-C2 )
132.1(8), C1-C2-C3 ) 126.0(8).
Sch em e 4
Sch em e 5
developed by our group,11 could be converted into
isolable Ni(0) π complexes (Scheme 4, Figure 3). Under
the same conditions that resulted in the formation of
the metallacycles from alkynyl enals described above,
treatment of bis-enone 15 with stoichiometric quantities
of Ni(COD)2 and bipyridine resulted in the formation
of complex 16. Although the substrate structure varies
(bis-enone vs alkynyl enal), it is noteworthy that the
ligand systems are identical within the two classes of
structures.
Dienes are yet another type of unsaturated subunit
that has been utilized in many classes of nickel-
catalyzed reactions;6,8c-f thus, we were interested in
probing the structure of Ni(0) diene complexes within
the same series of well-defined compounds described
above. Treatment of enone diene 17 with Ni(COD)2 and
bipyridine afforded Ni(0) adduct 18 (Scheme 5, Figure
4). As observed with bis-enones, the simple tetrahedral
Ni(0) bis π complex was observed. Compound 18 proved
to be very unstable, and bulk material was not obtained
in satisfactory purity to allow spectroscopic character-
ization.
F igu r e 2. X-ray crystal structure of metallacycle 14b.
Selected bond lengths (Å) and angles (deg): Ni-O )
1.8347(17), O-C11 ) 1.319(3), C11-C12 ) 1.324(4), C12-
C13 ) 1.493(4), C17-C18 ) 1.333(3), C18-Ni ) 1.897(2);
O-Ni-C18 ) 92.13(8), O-Ni-N1 ) 89.62(8), N1-Ni-N2
) 82.35(9), N2-Ni-C18 ) 95.87(9), Ni-O-C11 ) 127.5(2),
O-C11-C12 ) 130.6(3), C11-C12-C13 ) 125.8(3).
in the NMR spectra of nickel C-enolates, the carbon and
protons R to the carbonyl appeared at δ -8.0 to 13.6
(13C) and δ 0.75 to 1.88 (1H) compared with the analo-
gous signals for complex 14a , which appeared at δ 101
(13C) and δ 3.5 (1H). The corresponding bipyridine
complex 14b was also prepared, and its structural and
spectroscopic properties were directly analogous to those
of complex 14a (Scheme 3, Figure 2).
The isolation of complexes 14a and 14b provides
structural details about the metallacycle derived from
oxidative cyclization of an enone and alkyne, but the
features of the starting bis π complex itself (prior to
oxidative cyclization) must also be considered in formu-
lating a complete picture of the overall transformation.
Therefore, we were very pleased to observe that bis-
enones, which were also utilized in catalytic cyclizations
Rea ctivity Stu d ies
With considerable information in hand about struc-
ture, we next examined the reactivity of metallacycle
14a , derived from oxidative cyclization of enal-alkyne
(10) (a) Burkhardt, E. R.; Bergman, R. G.; Heathcock, C. H.
Organometallics 1990, 9, 30. (b) See ref 9d for additional discussion of
nickel enolates.
(11) (a) Savchenko, A. V.; Montgomery, J . J . Org. Chem. 1996, 61,
1562. (b) Montgomery, J .; Oblinger, E.; Savchenko, A. V. J . Am. Chem.
Soc. 1997, 119, 4911.

Combined Experimental and Computational Investigation
Organometallics, Vol. 23, No. 20, 2004 4639
Sch em e 7
F igu r e 4. X-ray crystal structure of complex 18. Selected
bond lengths (Å) and angles (deg): Ni-C18 ) 2.052(6), Ni-
C19 ) 1.977(5), Ni-C23 ) 2.016(6), Ni-C24 ) 2.063(6);
C19-Ni-C23 ) 97.7(3).
on the order of 1-2 h being required to achieve complete
conversion. Thus, the conversions of both 13a to 14a
and of 14a to 19a proceed too slowly to be involved in
the catalytic conversion of 13a to 19a . Computational
investigation of a model for the 13a to 14a conversion
provided further insight into this transformation and
suggested an enthalpic basis for these empirical obser-
vations (vide infra).
Sch em e 6
To rigorously establish the conclusions from these
qualitative experimental observations, we examined the
catalytic cyclization of 13b with dimethylzinc using
metallacycle 14a (20 mol %) as the catalyst (Scheme 7).
If metallacycle 14a is a kinetically competent species,
then a catalytic cyclization that proceeds to completion
should generate a 1:5 ratio of 19a and 19b. However,
the catalytic cyclization of 13b with dimethylzinc using
catalyst 14a (20 mol %), in a 5 min reaction, cleanly
affords products 19b and 19a in a 98:2 ratio. Bicycle
20, derived from protonation of metallacycle 14a upon
workup, was isolated in 82% yield based on the original
loading of metallacycle 14a . Thus, a small portion
(<10%) of the metallacycle 14a is converted to product
19a while generating a new nickel species that catalyzes
the formation of 19b from 13b much more rapidly than
14a is converted to 19a . The majority of metallacycle
14a persists until the reaction is quenched, which
directly converts 14a into bicycle 20.
To provide further insight into differences between
the stoichiometric transformations involving metalla-
cycle 14a with the overall catalytic conversion of alkynyl
enal 13a into 19a , we examined both sets of transfor-
mations with a chiral nickel species (Scheme 8). Given
the structural similarity of bis-oxazolines to tmeda and
bipyridine, which stabilize the nickel metallacycles in
question, we examined asymmetric transformations
involving ligand 23. Catalytic cyclizations involving Ni-
(COD)2 and ligand 23 (1:1, 10 mol %), were quite
efficient, with good yields of product 19a being obtained
in 77% yield after <5 min at room temperature.
However, the reactions proceeded essentially without
enantioselection, and measured ee’s were <5%. Next,
the stoichiometric conversion of alkynyl enal 13a to the
metallacycle upon treatment with 1.1 equiv of the 1:1
adduct of Ni(COD)2 and ligand 23 was carried out, and
quenching with methanol afforded bicyclooctanol 20 in
33% yield in 72:28 er. The stereochemistry-determining
steps of the potentially similar catalytic production of
19a and stoichiometric production of 20 clearly do not
substrate 13a . Complex 14a was observed to demon-
strate reactivity that closely parallels the nickel-
catalyzed chemistry of alkynyl enal 13a . For instance,
treatment of complex 14a with dimethylzinc (generated
in situ from MeLi and ZnCl2) affords a 71% isolated
yield of aldehyde 19a , in direct analogy to the catalytic
alkynyl enone/organozinc couplings that we extensively
developed (Scheme 6). In addition, the stoichiometric
conversion of 14a to 20-22 via nickel enolate protona-
tion with water or alkylation with either methyl iodide
or benzaldehyde proceeded with yield and selectivity
similar to those of the in situ procedure recently
disclosed from our laboratories.9b,d
The behavior of nickel metallacycle 14a described
above clearly demonstrates the chemical competence of
nickel metallacycles in alkylative cyclizations involving
dimethylzinc, but the results tell nothing about the
kinetic competence of this species in catalytic cycliza-
tions. The catalytic cyclization of alkynyl enal 13a with
Ni(COD)2/tmeda (10 mol %, 1:1) and commercial di-
methylzinc (3 equiv) at 0 °C in THF (0.1 M in alkynyl
enal) leads to clean production of the expected compound
19a in a very fast reaction, with complete conversion
in <5 min at 0 °C. In contrast, treatment of alkynyl enal
13a with a stoichiometric quantity of Ni(COD)2/tmeda
(1:1) produces metallacycle 14a in a much slower
process, with reaction times on the order of 1-2 h being
required to achieve complete conversion. Similarly,
treatment of metallacycle 14a with 3.0 equiv of dimeth-
ylzinc produces compound 19a more slowly than cata-
lytic cyclizations of 13a proceed, with reaction times also

4640 Organometallics, Vol. 23, No. 20, 2004
Hratchian et al.
Sch em e 8
26 are too slow to be involved in the catalytic process.
In addition, the structure of the main-group organome-
tallic was observed to have an effect on not only the
regiochemistry of alkyne insertions but also the overall
constitution of the products in various instances involv-
ing related reaction classes.8e,13 It therefore became
apparent that some modification of the proposed mech-
anism must be formulated.
Since the above experimental studies clearly demon-
strated that the presence of organozincs in some way
alters the reaction progression and rate in the interac-
tion of Ni(0) with 13a , we opted to employ a computa-
tional approach to provide insight into the precise role
of the organozinc participation. Electronic structure
calculations were carried out using the GAUSSIAN
suite of programs.14 All wave functions are closed-shell
singlets.15 The B3LYP density functional method and
the 6-31G(d) basis set were employed.16 Equilibrium and
transition state structures were optimized using stan-
dard methods17 and were verified by computing analytic
second derivatives. Transition states were further veri-
fied by visualization of transition vectors and/or reaction
path following calculations.18
Sch em e 9
To computationally evaluate modified mechanisms,
we first evaluated our originally proposed mechanism
described in Scheme 9. As a first step in our computa-
tional studies, the ground-state conformation of com-
pound 27 possessing an ethylenediamine ligand was
minimized to verify that there was an excellent agree-
(13) For example, compare ref 13a-c: (a) Oblinger, E.; Montgomery,
J . J . Am. Chem. Soc. 1997, 119, 9065. (b) Huang, W.-S.; Chan, J .;
J amison, T. F. Org. Lett. 2000, 2, 4221. (c) Ni, Y.; Amarasinghe, K. K.
D.; Montgomery, J . Org. Lett. 2002, 4, 1743. Also see: (d) Lozanov,
M.; Montgomery, J . J . Am. Chem. Soc. 2002, 124, 2106.
proceed in identical fashion. Interpretations of these
intriguing pieces of data regarding structure, reactivity,
reaction rates, and enantioselection will be discussed
below.
(14) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J . R.; Montgomery, J . A., J r.; Vreven, T.;
Kudin, K. N.; Burant, J . C.; Millam, J . M.; Iyengar, S. S.; Tomasi, J .;
Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson,
G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.;
Hasegawa, J .; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai,
H.; Klene, M.; Li, X.; Knox, J . E.; Hratchian, H. P.; Cross, J . B.; Adamo,
C.; J aramillo, J .; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin,
A. J .; Cammi, R.; Pomelli, C.; Ochterski, J . W.; Ayala, P. Y.; Morokuma,
K.; Voth, G. A.; Salvador, P.; Dannenberg, J . J .; Zakrzewski, V. G.;
Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J . B.; Ortiz, J . V.; Cui,
Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J .; Stefanov, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J .;
Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challa-
combe, M.; Gill, P. M. W.; J ohnson, B.; Chen, W.; Wong, M. W.;
Gonzalez, C.; Pople, J . A.; GAUSSIAN 03; Gaussian, Inc.: Pittsburgh,
PA, 2003.
(15) All transition states were tested for SCF stability. Only TS-30
and TS-34 were found to have minor instabilities, and recalculating
these structures as open-shell singlets lowered the energy by only 0.336
and 0.014 kcal/mol and yielded values of S2 of 0.212 and 0.046,
respectively. Further, the eigenvalues of the orbital rotation Hessian
are -0.009 568 9 and -0.001 996 1 au, respectively. This indicates that
none of the transition states have significant multireference character
at this level of theory.
(16) (a) Becke, A. D. Phys. Rev. A 1988, 38, 3098. (b) Becke. A. D. J .
Chem. Phys. 1993, 98, 5648. (c) Lee, T.; Yang, W. T.; Parr, R. G. Phys.
Rev. B 1988, 37, 785. (d) Ditchfield, R.; Hehre, W. J .; Pople, J . A. J .
Chem. Phys. 1971, 54, 724. (e) Gordon, M. S. Chem. Phys. Lett. 1980,
76, 163. (f) Hariharan, P. C.; Pople, J . A. Theor. Chim. Acta 1973, 28,
213. (g) Hariharan, P. C.; Pople, J . A. Mol. Phys. 1974, 27, 209. (h)
Hehre, W. J .; Ditchfield, R.; Pople, J . A. J . Chem. Phys. 1972, 56, 225.
(17) (a) Schlegel, H. B. J . Comput. Chem. 1982, 3, 214. (b) Peng, C.;
Schlegel, H. B. Isr. J . Chem. 1993, 33, 449. (c) Peng, C.; Ayala, P. Y.;
Schlegel, H. B.; Frisch, M. J . J . Comput. Chem. 1996, 17, 49.
(18) (a) Gonzalez, C.; Schlegel, H. B. J . Chem. Phys. 1989, 90, 2154.
(b) Gonzalez, C.; Schlegel, H. B. J . Phys. Chem. 1990, 94, 5523. (c)
Hratchian, H. P.; Schlegel, H. B. J . Phys. Chem. A 2002, 106, 165. (d)
Hratchian, H. P.; Schlegel, H. B. J . Chem. Phys. 2004, 120, 9918-
9924. (e) Hratchian, H. P.; Schlegel, H. B. J . Chem. Theor. Comput.,
in press.
Com p u ta tion a l An a lysis of th e Or ga n ozin c
In volvem en t
Our original interpretation of the mechanism of
nickel-catalyzed cyclizations of alkynyl enals and orga-
nozincs followed well-precedented chemical steps (Scheme
9). Complexation of Ni(0) to the alkynyl enone would
generate tetrahedral complex 24. Oxidative cyclization
of 24 would result in the formation of metallacycle 14
concomitant with geometric reorganization at nickel
from tetrahedral to square planar as the oxidation state
progresses from Ni(0) to Ni(II). Transmetalation of
dimethylzinc would result in the formation of interme-
diate 25, which produces the zinc enolate 26 of the
observed product 19a upon reductive elimination. In the
course of these studies, complex 14 was directly isolated
and characterized as the tmeda and bipyridine adducts,
and two chemical models for bis π complex 24 were
obtained by replacing the alkyne unit with an enone or
diene.12 However, the experiments presented above
(Scheme 7 and accompanying discussion) are inconsis-
tent with this mechanistic picture, since both the
conversion of 13a to 14a and the conversion of 14a to
(12) (a) For examples of Ni(0) complexes that possess and alkene
and an alkyne unit, see: Po¨rschke, K.-R. J . Am. Chem. Soc. 1989, 111,
5691. (b) For an example of a nickel(II) metallacycle derived from an
alkene and an alkyne unit, see ref 1d.

Combined Experimental and Computational Investigation
Organometallics, Vol. 23, No. 20, 2004 4641
F igu r e 6. Energy changes of the 29 to 28 reaction at the
B3LYP/6-31G(d) level. Values in parentheses are energies
relative to the reactant in kcal/mol.
Ta ble 1. Selected Bon d Len gth s (Å) a n d An gles
(d eg) of 14a , 27, a n d 28
14a
27
28
Ni-O
Ni-C
Ni-N1
Ni-N2
1.852(6)
1.897(7)
1.982(6)
2.077(7)
1.8138
1.8742
1.9314
2.0303
1.8157
1.8744
1.9301
2.0254
O-Ni-C
91.3(3)
94.3(3)
85.6(3)
88.9(3)
98.10
91.79
84.61
85.51
97.85
92.09
84.81
85.28
C-Ni-N1
N1-Ni-N2
N2-Ni-O
motif. This intermediate was 13.3 kcal/mol lower in
energy than the starting π complex 29 and, thus,
slightly higher in energy (by 0.6 kcal/mol) than the
rearranged O-enolate 28. The transition-state structure
and energies for the conversion of 29 to 30 and for 30
to 28 were determined. The barrier for the oxidative
cyclization (29 to 30) was 30.8 kcal/mol, and the barrier
for the C-enolate to O-enolate tautomerization (30 to
28) was 19.8 kcal/mol.19 As expected, the transition state
for the 29 to 30 conversion involves formation of the
C-C bond and geometric reorganization of the transi-
tion-metal center from tetrahedral to square planar. The
geometry at nickel of TS-30 is slightly distorted tetra-
hedral. A stability test of the closed-shell singlet wave
function indicated a slight RHF/UHF instability.15 This
suggests that the oxidation of Ni(0) in the 29 to 30
conversion occurs early in the reaction and occurs more
quickly than the geometric rearrangement. The transi-
tion state for the 30 to 28 conversion adopts an enolate
binding motif that resembles an η3-enolate structure as
the tautomerization and ring expansion occurs. Most
features of this overall conversion are consistent with
expectations, but the energy barrier for the initial
oxidative cyclization is very high for an efficient catalytic
process. This high barrier for oxidative cyclization is
consistent with the experimental observation that the
F igu r e 5. Optimized geometries for 27 and minima and
transition states involved in the formation of 28 from 29.
ment between the solid-state structure of 14a , as
determined by X-ray crystallography (Figure 1), and the
computationally derived conformation of structure 27
(Figure 5). The conformation of the simpler metallacycle
28 (Figure 5) derived from crotonaldehyde and propyne
was then determined computationally, and all three
structures 14a , 27, and 28 were found to provide good
agreement in the structural characteristics within the
metallacycle framework (Table 1). The crystallographi-
cally characterized bis-enone π complex structure 16
(Figure 3) was then modified to include the relevant enal
and alkyne functionalities and the ethylenediamine
ligand motif, and minimization of the energy of that
structure then provided the ground-state complex 29
that would lead to metallacycle 28 upon oxidative
cyclization (Figure 5).
Comparison of the relative energies of the two struc-
tures 28 and 29 revealed that metallacycle 28 was more
stable by 13.9 kcal/mol (Figure 6). In searching for a
transition state for this conversion, intermediate 30 was
found on the reaction path that possessed a five-
membered metallacycle with an η1 C-enolate structural
(19) For a recent discussion of C- and O-tautomers of nickel enolates,
see: Campora, J .; Maya, C. M.; Palma, P.; Carmona, E.; Gutie´rrez-
Puebla, E.; Ruiz, C. J . Am. Chem. Soc. 2003, 125, 1482.

4642 Organometallics, Vol. 23, No. 20, 2004
Hratchian et al.
F igu r e 7. Energy changes of the 31a to 34 reaction at
the B3LYP/6-31G(d) level. Values in parentheses are
energies relative to the reactant in kcal/mol.
F igu r e 8. Optimized geometries for the structures in-
volved in the formation of 34 from 31a .
Ch a r t 1
Ch a r t 2. Cr ysta llogr a p h ica lly Ch a r a cter ized
Ad d u cts (P o1r sch k e a n d Wilk e)
stoichiometric oxidative cyclization proceeds slowly in
the absence of organozinc. The basic features of the
proposed mechanism may be correct, but if metallacycle
formation is involved in the process, then the organozinc
must be involved in a rate acceleration of the metalla-
cycle-forming process.
We consider three roles to be most likely for the
organozinc: Lewis acidic activation of the enal carbonyl
(31), Lewis basic activation of a Ni(0) species via three-
center-two-electron bridging of the zinc-methyl bond
with nickel (32a ), or a combination of the two effects
(32b) (Chart 1).20
Lewis acid activation of the enal carbonyl was first
considered (Figures 7 and 8). In our computational
analysis, Lewis acidic activation of the carbonyl oxygen
results in the C-enolate metallacycle 34, which is formed
with an activation barrier of 26.3 kcal/mol. Similar to
the 29 to 30 conversion, the oxidation of Ni(0) occurs
more rapidly than the geometric rearrangement of the
transition-metal center in the 31a to 34 conversion.15
The five-membered C-enolate metallacycle 34 is more
stable than the ground-state π complex 31a by 15.8 kcal/
mol. An absolute comparison of the ground-state ener-
gies of 29 and 31a or of 30 and 34 is not meaningful,
since the structures are not constitutional isomers.
However, the relatively large ∆Hq value for the orga-
nozinc-promoted pathway suggests that a simple Lewis
acidic interaction of the organozinc reagent does not
satisfactorily explain the efficient catalysis observed.
A Lewis basic interaction of the organozinc with Ni(0)
was considered next. In searching for examples of
various metal alkyls and metal hydrides that interact
with Ni(0) complexes, work from Po¨rschke and Wilke
emerged as the most relevant structures that directly
document such an interaction (Chart 2).1a,21 Studies
from those laboratories elegantly demonstrated that
Ni(0) serves as a Lewis acid in the presence of metal
alkyls and metal hydrides. In particular, two relevant
crystal structures have been reported: one (35) docu-
ments the direct coordination of tmeda-ligated dimeth-
ylmagnesium with a Ni(0) bis(ethylene) complex, and
the second (36) documents the direct coordination of
quinuclidine-ligated dimethylaluminum hydride with a
Ni(0) cyclododecatriene complex. We envision that these
coordination modes may serve as excellent models for
the manner in which organozincs, organosilanes, and
other reducing agents may interact with the Ni(0) π
complexes that are relevant to the catalytic processes
developed in our laboratory. Such an interaction could
explain the role of the organozinc in controlling the rate,
(20) For discussion of interactions of metal hydrides or metal alkyls
with nickel in catalytic processes, see: (a) Lautens, M.; Ma, S.; Chui,
P. J . Am. Chem. Soc. 1997, 119, 6478. (b) Lautens, M.; Rovis, T. J .
Am. Chem. Soc. 1997, 119, 11090. (c) Terao, J .; Oda, A.; Ikumi, A.;
Nakamura, A.; Kuniyasu, H.; Kambe, N. Angew. Chem., Int. Ed. 2003,
42, 3412. (d) Subburaj, K.; Montgomery, J . J . Am. Chem. Soc. 2003,
125, 11210. (e) Mahandru, G. M.; Liu, G.; Montgomery, J . J . Am. Chem.
Soc. 2004, 126, 3698.
(21) (a) Kaschube, W.; Po¨rschke, K.-R.; Angermund, K.; Kru¨ger, C.;
Wilke, G. Chem. Ber. 1988, 121, 1921. (b) Po¨rschke, K.-R.; Kleimann,
W.; Tsay, Y.-H.; Kru¨ger, C.; Wilke, G. Chem. Ber. 1990, 123, 1267. (c)
J onas, K.; Po¨rschke, K.-R.; Kru¨ger, C.; Tsay, Y.-H. Angew. Chem., Int.
Ed. Engl. 1976, 15, 621. (d) Po¨rschke, K.-R.; J onas, K.; Wilke, G. Chem.
Ber. 1988, 121, 1913.

Combined Experimental and Computational Investigation
Organometallics, Vol. 23, No. 20, 2004 4643
F igu r e 9. Energy changes of the 32a to 38a reaction and the 32b to 38b reaction at the B3LYP/6-31G(d) level. Values
in parentheses are energies relative to 32a in kcal/mol.
F igu r e 10. Optimized geometries for the structures involved in the formation of 38a from 32a , 38b from 32b, and 38c
from 32b.
regiochemistry, and constitution of products obtained
in catalytic cyclizations and additions.
activation barrier for the oxidative cyclization of 32a is
quite low (∆Hq ) 10.4 kcal/mol). In contrast, adduct 32b,
whose ground-state energy is 5.7 kcal/mol lower than
that of adduct 32a , undergoes oxidative cyclization to
a metallacycle with a higher barrier of ∆Hq ) 20.5 kcal/
mol. Thus, the energy of TS-37a is lower than that of
TS-37b by 4.4 kcal/mol.
Lewis basic activation of a Ni(0) species via three-
center-two-electron bridging of the zinc-methyl bond
with nickel can proceed via the two isomeric ground-
state adducts 32a and 32b (Figures 9 and 10), with
isomer 32b possessing an additional stabilizing Zn-O
interaction not possible in 32a . Adduct 32a is the least
stable of these by 5.7 kcal/mol, but the resulting
Most interestingly, an alternate rearrangement path-
way that involves 32b provides the lowest barrier for

4644 Organometallics, Vol. 23, No. 20, 2004
Hratchian et al.
F igu r e 11. Energy changes of the 32b to 38c reaction at
the B3LYP/6-31G(d) level. Values in parentheses are
energies relative to 32a in kcal/mol.
F igu r e 12. Key orbital interactions in 32b: (a) unoccupied
Zn-C σ* orbital; (b) occupied Ni dz orbital; (c) oxygen lone-
pair orbital involved in donation to Zn-centered orbital.
2
the overall transformation (Figures 10 and 11). Starting
from adduct 32b, a formal oxidative addition of Ni(0)
to the Lewis acid activated enone proceeds with a very
low barrier of 2.5 kcal/mol to produce the bicyclic π-allyl
intermediate 33,22 which itself is 2.5 kcal/mol less stable
than adduct 32b. In contrast to the 20.5 kcal/mol barrier
observed for the direct oxidative cyclization of 32b,
adduct 33 undergoes cyclization to metallacycle 38c
with a barrier of only 6.2 kcal/mol, making the overall
barrier for the conversion of 32b to 33 to metallacycle
38c significantly lower than any other pathway exam-
ined. The conversion of 33 to 38c is accompanied by a
weakening of the bridging zinc-carbon and zinc-nickel
interactions. Therefore, structure 38c is best repre-
sented as a hybrid of the two resonance extremes
depicted (Figure 11).
32b involves σ-donation of Zn-C bonding electrons to
the unoccupied nickel 4s orbital.24 However, an unusual
bonding interaction predicted by theory is the significant
2
back-bonding of the filled Ni 3dz orbital with the σ*
orbital of the nonbridging Zn-C bond (Figure 12). It is
this latter interaction that causes the short interatomic
distance between Ni and Zn (2.3 Å).25 For both 32a and
32b, the stabilization energies arising from these two
interactions between the Me2Zn and Ni fragments are
roughly equivalent (within 2.5 kcal/mol). 32b is further
stabilized by an interaction between the enone and
Me2Zn fragments. This interaction, which is not present
in 32a , is characterized by donation from a lone-pair
orbital on O to the same Zn-nonbridging Me σ* orbital
involved in back-donation from Ni.
The three-membered ring observed in the character-
ized MgMe2 adduct of Ni0 and the computationally
predicted ZnMe2 adducts of Ni0 involve very unusual
interactions. We therefore employed the natural bond
orbital (NBO) analysis of Weinhold and co-workers to
examine the bonding in structures 32a and 32b.23 We
have specifically made use of the second-order perturba-
tion analysis of the canonical NBOs to determine the
stabilization energies of NBOs due to interactions
between fragment orbitals. Each ligand on the nickel
has been defined as a separate fragment, and the Ni
atom is defined as a fourth fragment. Not surprisingly,
the key interaction of the bridging alkyl in both 32a and
(23) (a) Carpenter, J . E.; Weinhold, F. J . Mol. Struct. (THEOCHEM)
1988, 46, 41. (b) Carpenter, J . E. Ph.D. Thesis, University of Wisconsin,
1987. (c) Foster, J . P.; Weinhold, F. J . Am. Chem. Soc. 1980, 102, 7211.
(d) Reed, A. E.; Weinhold, F. J . Chem. Phys. 1983, 78, 4066. (e) Reed,
A. E.; Weinhold, F. J . Chem. Phys. 1985, 83, 1736. (f) Reed, A. E.;
Weinstock, R. B.; Weinhold, F. J . Chem. Phys. 1985, 83, 735. (g) Reed,
A. E.; Curtiss, L. A.; Weinhold, F. Chem. Rev. 1988, 88, 899. (h)
Weinhold, F.; Carpenter, J . E. The Structure of Small Molecules and
Ions, Plenum: Amsterdam, 1988, p 227.
(24) For an analysis of molecular orbitals in dimethylzinc, see: Mori,
S.; Hirai, A.; Nakamura, M.; Nakamura, E. Tetrahedron 2000, 56,
2805.
(25) (a) For an example of a Zn-Ni bond, see: Budzelaar, P. H. M.;
Boersma, J .; van der Kerk, G. J . M. Organometallics 1985, 4, 680.
(b) For an example of a Pt-Cd bond that involves donation from a
2
filled dz orbital to a vacant orbital on Cd, see: Yamaguchi, T.;
Yamazaki, F.; Ito, T. J . Am. Chem. Soc. 1999, 121, 7405. (c) For a
general discussion of transition-metal/main-group-metal bonding,
see: Frenking, G. F. J . Organomet. Chem. 2001, 635, 9. (d) For a
general discussion of the influence of back-bonding interactions on d10
metal reactivity, see: Stahl, S. S.; Thorman, J . L.; de Silva, N.; Guzei,
I. A.; Clark, R. W. J . Am. Chem. Soc. 2003, 125, 12.
(22) (a) J ohnson, J . R.; Tully, P. S.; Mackenzie, P. B.; Sabat, M. J .
Am. Chem. Soc. 1991, 113, 6172. (b) Grisso, B. A.; J ohnson, J . R.;
Mackenzie, P. B. J . Am. Chem. Soc. 1992, 114, 5160. (c) Ogoshi, S.;
Yoshida, T.; Nishida, T.; Morita, M.; Kurosawa, H. J . Am. Chem. Soc.
2001, 123, 1944.

Combined Experimental and Computational Investigation
Organometallics, Vol. 23, No. 20, 2004 4645
Sch em e 10
Given these bonding characteristics, the question then
arises regarding the origin of differences among TS-37a ,
TS-37b, and TS-33. Our analysis suggests that TS-37a
is lower in energy than TS-37b because the weak zinc-
oxygen interaction present in 32b is broken in TS-37b
as the nickel enolate adopts an η3 interaction. However,
the weak zinc-oxygen interaction of 32b is further
strengthened in TS-33, and that favorable interaction
leads to the overall low barrier for the 32b to 38c
conversion.
Discu ssion a n d Im p lica tion s
This mechanism involving a dual Lewis basic/Lewis
acidic role of the organozinc (Figure 11) possesses a
number of attractive features. It involves the core idea
of metallacycle involvement, which explains many of the
synthetic results, and it is consistent with the qualita-
tive kinetic observations described above. The important
experiment described in Scheme 7, which demonstrated
that metallacycle 14a is not kinetically competent in the
catalytic conversion of alkynyl enal 13b to product 19b,
may be rationalized by a mechanism that involves
formation of an active Ni(0)/ZnMe2 adduct. The trace
quantity of product 19a obtained is derived from
conversion of a small portion of metallacycle 14a into
product 19a . Interaction of dimethylzinc with the
nickel(0) species generated by production of compound
19a then generates a more active catalyst that rapidly
converts substrate 13b into product 19b, perhaps by the
mechanism depicted in Figure 11. The majority of
metallacycle 14a is unchanged during the short reaction
time (5 min) and is converted upon workup into bicycle
20.
Furthermore, as the rates of many metal-catalyzed
cross-couplings are restricted by a rate-determining
transmetalation event,26 the proposed prior coordination
of the organozinc avoids this potential complexity, since
a unimolecular rearrangement is required for metalla-
cycle 38c to be converted to product. Although the final
stages of the overall catalytic process were not modeled
in this study, it is interesting to note that intermediate
38c is ideally positioned to undergo the required carbon-
carbon bond-forming reductive elimination that would
allow product formation. Alternatively, structure 34,
which is formed by the Lewis acidic role of the orga-
nozinc (Figures 7 and 8), would likely require a poten-
tially high barrier ligand dissociation during the Zn to
Ni transmetalation event. Furthermore, we reiterate
that Ni(COD)2 itself in the absence of diamines is the
typical formulation for efficient catalytic cyclizations,
and the diamine formulations were investigated simply
to allow study of the kinetic competence of the tmeda
metallacycle that was isolated. Thus, an overall mech-
anism that incorporates the experimental observations
reported herein and that embodies the computationally
predicted interactions of dimethylzinc and Ni(0) may be
formulated as shown (Scheme 10).
the absence of organozincs likely proceeds with biden-
tate coordination of the bis(oxazoline) in a stereochem-
istry-determining oxidative cyclization. Methanol quench-
ing results in liberation of the bicyclooctenol product 20
with conservation of enantioselectivity. If the orga-
nozinc-mediated cyclization to give product 19a proceeds
via the same metallacycle, then the enantioselectivities
for the production of compounds 19a and 20 would likely
be similar. However, if the organozinc serves to seques-
ter the bis(oxazoline), then an equilibrium concentration
of a more reactive achiral catalyst, which involves an
association of nickel with ZnMe2 (as depicted in Figure
11 and Scheme 10), rather than with the bis(oxazoline),
could be responsible for the production of racemic 19d .
Another potential explanation is that 19a is produced
from a metallacycle generated under kinetic control with
poor enantioselection, whereas 20 is produced from the
metallacycle generated stoichiometrically under ther-
modynamic conditions with better (72:28) enantioselec-
tion,27 although we favor the former explanation. In-
deed, the ground-state interactions depicted in the
crystallographically characterized adduct of dimethyl-
magnesium, tmeda, and Ni(0) from Po¨rschke and Wilke
(Chart 2) depicts exactly this type of interaction wherein
direct association of Ni(0) with the diamine is clearly
not involved.
We believe that the above mechanistic suggestions
could have implications in many classes of related
transformations. For instance, one previously unex-
plained observation from our group and from J amison
is the diverging selectivities for ethyl vs hydrogen atom
incorporation in nickel-catalyzed cyclizations and cou-
plings involving either diethylzinc or triethylborane.8g,13
If the extent of bridging between nickel and either zinc
or boron is significantly different, then significant
differences in the relative rates of processes involving
selectivity-determining reductive elimination and â-hy-
dride elimination from a bridged ethyl species would not
be at all surprising. Another unexplained feature of
aldehyde-alkyne couplings from our earlier work is the
regiochemical reversal of alkyne insertion in intermo-
lecular processes involving either alkenylzirconocenes
or alkylzincs as the reducing agent.13 If the oxidative
The divergent enantioselectivity results (Scheme 8)
may also be readily rationalized by this mechanism. The
[3 + 2] reductive cycloaddition to form product 20 in
(27) We are unaware of an exact precedent for this possibility.
However, very different kinetic and thermodynamic cis/trans selectivi-
ties in the zirconocene-catalyzed cyclomagnesiation of dienes is well
established: (a) Knight, K. S.; Wang, D.; Waymouth, R. M.; Ziller, J .
J . Am. Chem. Soc. 1994, 116, 1845. (b) Taber, D. F.; Louey, J . P.; Wang,
Y.; Nugent, W. A.; Dixon, D. A.; Harlow, R. L. J . Am. Chem. Soc. 1994,
116, 9457.
(26) (a) Amatore, C.; Bahsoun, A. A.; J utand, A.; Meyer, G.; Ntepe,
A. N.; Ricard, L. J . Am. Chem. Soc. 2003, 125, 4212. (b) Louie, J .;
Hartwig, J . F. J . Am. Chem. Soc. 1995, 117, 11598. (c) Farina, V.;
Krishnan, B. J . Am. Chem. Soc. 1991, 113, 9585.

4646 Organometallics, Vol. 23, No. 20, 2004
Hratchian et al.
cyclization involves a Ni(0) species complexed to the
organozinc or alkenylzirconium, then the dramatically
different steric properties of these two types of reagents
would be expected to play an important role in alkyne
regioselection. We do not mean to imply that ligand
sequestration routinely occurs in reactions of this
general class. Important ligand effects have been docu-
mented, and various enantioselective chiral ligand-
based transformations have been reported.8g,28 However,
there may be instances where organozincs can sequester
certain ligand classes, or their interaction with Ni(0)
catalysts may occur in concert with more classical
metal-ligand interactions such as phosphine-nickel
dative bonding.
In an even broader context, we suggest that the
interaction of main-group transmetalating agents and
late-metal catalysts may play a role in the oxidative
addition step of many classes of cross-coupling pro-
cesses. Elegant studies from Amatore and J utand have
clearly demonstrated that chloride association with
Pd(0) generates an active catalyst in oxidative addition
processes.29 A similar interaction of transmetalating
agents with late-metal cross-coupling catalysts via the
interaction modes described herein may, in some in-
stances, be involved in catalytic mechanisms as well.30
Ack n ow led gm en t. J .M. thanks the National Sci-
ence Foundation (Grant No. CHE-0405800), J ohnson
and J ohnson (Focused Giving Grant), the Camille and
Henry Dreyfus Foundation (Teacher Scholar Award),
and the American Chemical Society (Arthur C. Cope
Scholar Award) for support, H.B.S. thanks the National
Science Foundation (Grant No. CHE 0131157) and the
National Computational Science Alliance (Grant No.
CHE980042N) for support, V.M.G.-G. thanks the CON-
ACYT-Me´xico for a postdoctoral fellowship, and H.P.H.
thanks the Institute for Scientific Computing at Wayne
State University for support provided by an NSF-IGERT
fellowship.
Su p p or tin g In for m a tion Ava ila ble: Text giving full
experimental details, figures giving 1H NMR spectra of all new
compounds, and tables giving X-ray crystallographic data and
SCF energies and gas-phase Cartesian coordinates for the
structures studied computationally; crystallographic data are
also given as CIF files. This material is available free of charge
via the Internet at http://pubs.acs.org.
OM049471A
(28) (a) Ikeda, S.; Cui, D.-M.; Sato, Y. J . Am. Chem. Soc. 1999, 121,
4712. (b) Sato, Y.; Saito, N.; Mori, M. J . Am. Chem. Soc. 2000, 122,
2371. (c) Montgomery, J .; Savchenko, A. V. J . Am. Chem. Soc. 1996,
118, 2099.
(29) (a) Amatore, C.; Azzabi, M.; J utand, A. J . Am. Chem. Soc. 1991,
113, 8375. (b) Amatore, C.; J utand, A.; Suarez, A. J . Am. Chem. Soc.
1993, 115, 9531. (c) Amatore, C.; J utand, A. Acc. Chem. Res. 2000, 33,
314. (d) Amatore, C.; Carre´, E.; J utand, A.; Tanaka, H.; Ren, Q.; Torii,
S. Chem. Eur. J . 1996, 2, 957.
(30) Although we believe that reaction acceleration by coordination
of a transmetalation agent prior to oxidative addition should be
considered in some cross-coupling processes, it is clearly not a general
phenomenon. For instance, prior coordination of an alkenylstannane
to Pd(AsPh3)2 in Stille couplings was demonstrated to have a decelerat-
ing effect: Amatore, C.; Bahsoun, A. A.; J utand, A.; Meyer, G.; Ntepe,
A. N.; Ricard, L. J . Am. Chem. Soc. 2003, 125, 4212.